In recent years, a monitoring system using multiple cameras mainly used for monitoring use has been actively researched and developed. Monitoring with multiple cameras requires fulfilling a first requirement to maintain a state where regions which become blind spots are reduced relative to an entire subject region to be monitored (a monitoring subject region), and further a second requirement to acquire detailed imaging information concerning objects within the subject monitoring region.
In a conventional device using multiple cameras, automatic adjustment of a position of an imaging region and an imaging range for each camera fulfills the two requirements. Typical conventional devices using multiple cameras are described in Patent Reference 1 and Patent Reference 2.
First, the conventional device described in Patent Reference 1 is explained. FIG. 1 shows a device described in Patent Reference 1, that automatically adjusts the camera imaging region.
In FIG. 1, in a camera detection device 9010, a camera 9011 and a reflecting mirror 9012 image a detection subject throughout a broad imaging region; a mobile object extractor 9013 extracts the detection subject from the imaged picture; and a position information extractor 9014 extracts position information of the detection subject, so the camera detection device 9010 acquires the position information of the detection subject throughout the broad detection region. In a camera determination device 9020, a camera controller 9022 controls the traverse angle, the depression angle and the zoom ratio of the camera 9021 based on the position information of the detection subject, and the camera determination device 9020 acquires detailed information of the detection subject in order to image an enlarged image of the detected subject.
FIG. 2 shows detection regions for the camera detection device 9010 and the camera determination device 9020. In FIG. 2, black circles indicate the installation position of the camera detection device 9010, respectively, wherein the camera detection device 9010 is a secured camera. Circles or hexagons indicate respective imaging regions for the camera detection device 9010. As shown in FIG. 2, artificially installing each camera detection device 9010 with regularity enables detection on a steady basis without any blind spot within the detection subject regions comprising the subject monitoring regions.
Next, a conventional device described in Patent Reference 2 is explained. FIG. 3 shows a device that automatically adjusts the camera imaging regions, as described in Patent Reference 2.
In FIG. 3, a moving object detection camera 9211 for imaging an object throughout a broad imaging region changes its own imaging region by a posture control unit 9212, and a monitoring camera 9221 for imaging an enlarged picture of the object changes its own imaging region by a posture control unit 9222. In an image processor 9240, the imaging region for each camera is determined from the position of the detection subject extracted from the picture imaged by the moving object detection camera 9211 and the imaging region for each camera based upon information pre-stored in a camera angle of memory view unit 9231 and a camera angle of memory view unit 9232.
The imaging region determination method of each camera, for use with such conventional device, is described hereafter. FIG. 4, FIG. 5 and FIG. 6 are diagrams explaining the imaging region determination method of each camera, and respectively show pictures divided into several block pictures, imaged by the moving object detection camera 9211. First, the imaging region of the moving object detection camera 9211 is determined as follows: when a detection subject exists in the shaded blocks in FIG. 4, the postures of the moving object detection cameras 9211 are changed to the directions indicated by the arrows marked in the blocks of FIG. 5 corresponding to the block positions shown in FIG. 4, and the imaging regions for the cameras are changed, respectively. The imaging region for the moving object detection camera 9211 corresponding to each block position is individually specified, and the information is pre-set in the camera angle of memory view unit 9231. Next, the imaging regions for the monitoring cameras 9221 are determined as follows: when a detection subject exists in the block positions shown in FIG. 6, the postures of the monitoring cameras 9221 are changed to imaging regions indicated by a broken line, and the imaging regions for the cameras are changed. The imaging region for the monitoring camera 9221 corresponding to each block position is manually specified, and the information is pre-set in the camera angle of memory view unit 9232.
Characteristics of conventional devices using multiple cameras are summarized hereafter.
First, the camera detection device 9010 in the conventional device shown in Patent Reference 1 and the moving object detection camera 9211 in the conventional device shown in Patent Reference 2 play a role in detecting a detection subject within a broad range region, and the camera determination device 9020 in the conventional device shown in Patent Reference 1 and the monitoring camera 9221 in the conventional device shown in Patent Reference 2 respectively play a role in acquiring detailed information of the detection subject, such as an enlarged picture of the detected subject. As described, in the prior art, the role is specified for each camera, and a camera playing one role fulfills the first requirement, and another camera playing the other role fulfills the second requirement (characteristic 1 of the prior art).
In addition, in the conventional device shown in Patent Reference 2, for example, as the imaging region for the moving object detection camera 9211 is changed to the detection region moved to the upper left direction shown in the upper left block in FIG. 5 corresponding to the state change where the monitoring subject is on the upper left block in FIG. 4, the imaging region for each camera is determined and adjusted based on the information in a table format that defines camera behavior corresponding to a pre-assumed and manually prepared state. (characteristic 2 of the prior art).
In addition, the conventional device shown in Patent Reference 1, as shown in FIG. 2, manually installs fixed cameras at regular positions, fulfills the first requirement (characteristic 3 of the prior art).    Patent Reference 1: Japanese Laid-Open Patent Application No. 3043925 (FIG. 1 and FIG. 6)    Patent Reference 2: Japanese Laid-Open Patent Application No. 3180730 (FIG. 1 and FIG. 7 to FIG. 9)